Derek Su

Replace the gcc memory allocation with efficient one https://github.com/longhorn/longhorn/issues/4136

Hello @iosifnicolae2, Could you elaborate `evict a volume from a disk` more? What's the purpose of the feature?

> > Hello @iosifnicolae2, > > Could you elaborate `evict a volume from a disk` more? > > What's the purpose of the feature? > > hi, sure > >...

> Do you think this kind of involuntary factor would be still a concern for data integrity or data corruption after we introduced the below fixes in 1.3.2? Do you...

> Add more range checks for the fields under spec/status. Note here. Related to https://github.com/longhorn/longhorn/pull/3812. We need to define the supported k8s version range.

The issue mentioned in https://github.com/longhorn/longhorn/issues/3125 is due to that `fsfreeze` freezes the mount point of the given path. Despite the [man page](https://linux.die.net/man/8/fsfreeze) stating the path is the pathname of the...

> Is this still going to proceed? @ChipWolf Yes. Any concern or disadvantage of the current `sync` way?

**Performance benchmark** - AWS c5d.xlarge with direct nvme disks - The device mapper doesn't introduce significant performance impact. - Longhorn volume with single replica (v1.5.1) ``` IOPS (Read/Write) Random: 15,277...

> > If one wants to take a snapshot of the volume, it can execute `dmsetup suspend` and the filesystem is automatically frozen ([code](https://elixir.bootlin.com/linux/latest/source/drivers/md/dm.c#L2693)). In addition to preventing the race...

> As I know, the suspended DM device should be on the top of the storage stack, according to the [architecture](https://github.com/longhorn/longhorn/issues/6001#issuecomment-1653238434), there would be another longhorn device on it? The...